tlf-ae-summary.Rmd

# AE summary {#aesummary}

Following [ICH E3 guidance](https://database.ich.org/sites/default/files/E3_Guideline.pdf),
we need to summarize which participants were included in each efficacy analysis in Section 12.2, Adverse Events (AEs).

```{r}
library(haven) # Read SAS data
library(dplyr) # Manipulate data
library(tidyr) # Manipulate data
library(r2rtf) # Reporting in RTF format
```

In this chapter, we illustrate how to summarize AE information in a study.

```{r, out.width = "100%", out.height = "400px", echo = FALSE, fig.align = "center"}
knitr::include_graphics("tlf/tlf_ae_summary.pdf")
```

The data used to summarize AE information is in `adsl` and `adae` datasets.

```{r}
adsl <- read_sas("adam_data/adsl.sas7bdat")
adae <- read_sas("adam_data/adae.sas7bdat")
```

We first summarize participants in population by treatment group.

```{r}
pop <- adsl %>%
  filter(SAFFL == "Y") %>%
  rename(TRTAN = TRT01AN) %>%
  count(TRTAN, name = "tot")

pop
```

We transform the data to simplify the analysis of each required AE criteria of interest.

- With one or more adverse events
- With drug-related adverse events
- With serious adverse events
- With serious drug-related adverse events
- Who died"

```{r}
tidy_ae <- adae %>%
  mutate(
    all = SAFFL == "Y",
    drug = AEREL %in% c("POSSIBLE", "PROBABLE"),
    ser = AESER == "Y",
    drug_ser = drug & ser,
    die = AEOUT == "FATAL"
  ) %>%
  select(USUBJID, TRTAN, all, drug, ser, drug_ser, die) %>%
  pivot_longer(cols = c(all, drug, ser, drug_ser, die))

tidy_ae %>% head(4)
```

We summarize the number and percentage of participants who meet each AE criteria.

```{r}
ana <- tidy_ae %>%
  filter(value == TRUE) %>%
  group_by(TRTAN, name) %>%
  summarise(n = n_distinct(USUBJID)) %>%
  left_join(pop, by = "TRTAN") %>%
  mutate(
    pct = fmt_num(n / tot * 100, digits = 1),
    n = fmt_num(n, digits = 0),
    pct = paste0("(", pct, ")")
  )

ana %>% head(4)
```

We prepare reporting-ready dataset for each AE criteria.

```{r}
t_ae <- ana %>%
  pivot_wider(
    id_cols = "name",
    names_from = TRTAN,
    values_from = c(n, pct),
    values_fill = list(
      n = "   0",
      pct = "(  0.0)"
    )
  )

t_ae <- t_ae %>%
  mutate(name = factor(
    name,
    c("all", "drug", "ser", "drug_ser", "die"),
    c(
      "With one or more adverse events",
      "With drug-related adverse events",
      "With serious adverse events",
      "With serious drug-related adverse events",
      "Who died"
    )
  )) %>%
  arrange(name)
```

We prepare reporting-ready dataset for the analysis population.

```{r}
t_pop <- pop %>%
  mutate(
    name = "Participants in population",
    tot = fmt_num(tot, digits = 0)
  ) %>%
  pivot_wider(
    id_cols = name,
    names_from = TRTAN,
    names_prefix = "n_",
    values_from = tot
  )

t_pop
```

The final report data is saved in `tbl_ae_summary`.

```{r}
tbl_ae_summary <- bind_rows(t_pop, t_ae) %>%
  select(name, ends_with("_0"), ends_with("_54"), ends_with("_81"))

tbl_ae_summary
```

We start to define the format of the output.

```{r}
tbl_ae_summary %>%
  rtf_title(
    "Analysis of Adverse Event Summary",
    "(Safety Analysis Population)"
  ) %>%
  rtf_colheader(" | Placebo | Xanomeline Low Dose| Xanomeline High Dose",
    col_rel_width = c(3.5, rep(2, 3))
  ) %>%
  rtf_colheader(" | n | (%) | n | (%) | n | (%)",
    col_rel_width = c(3.5, rep(c(0.7, 1.3), 3)),
    border_top = c("", rep("single", 6)),
    border_left = c("single", rep(c("single", ""), 3))
  ) %>%
  rtf_body(
    col_rel_width = c(3.5, rep(c(0.7, 1.3), 3)),
    text_justification = c("l", rep("c", 6)),
    border_left = c("single", rep(c("single", ""), 3))
  ) %>%
  rtf_footnote("Every subject is counted a single time for each applicable row and column.") %>%
  rtf_encode() %>%
  write_rtf("tlf/tlf_ae_summary.rtf")
```

```{r, out.width = "100%", out.height = "400px", echo = FALSE, fig.align = "center"}
knitr::include_graphics("tlf/tlf_ae_summary.pdf")
```

The procedure to generate an AE summary table can be summarized as follow:

- Step 1: Read data into R, i.e., `adae` and `adsl`.
- Step 2: Summarize participants in population by treatment group, and we name the dataset as `t_pop`.
- Step 3: Summarize participants in population by required AE criteria of interest,
  and we name the dataset as `t_ae`.
- Step 4: Rowly bind `t_pop` and `t_ae` and format it by `r2rtf`.